28 August 2010. Researchers have mediated a merger of sorts between two distinct neurodegenerative conditions—amyotrophic lateral sclerosis and spinocerebellar ataxia (SCA)—and the proteins that cause them. Both, it seems, involve TDP-43, first associated with ALS, and ataxin-2, which is related to SCA. Which disease is most likely seems to come down to the length of a polyglutamine stretch in ataxin-2. In the August 26 Nature, scientists from the University of Pennsylvania in Philadelphia report that the mid-length polyglutamine expansions in ataxin-2 increase a person’s risk for ALS. It is well known that longer stretches lead to SCA type 2, while short-length repeats are harmless.
The work was a UPenn multi-lab collaboration. “We went all the way from yeast to finding a new genetic risk factor for ALS,” said senior author Aaron Gitler. “It really underscores the power of these simple models.” Gitler’s lab started the study in yeast and later recruited co-senior author Nancy Bonini and colleagues for their fruit fly skills and John Trojanowski and Virginia Lee for their collections of human tissue samples. Joint first authors were Andrew Elden, Hyung-Jun Kim, Michael Hart, and Alice Chen-Plotkin.
Both TDP-43 and ataxin-2 are RNA-binding proteins of somewhat hazy function. TDP-43 pathology, in which the normally nuclear protein moves to the cytoplasm and forms aggregates, is common in ALS and also occurs in frontotemporal lobar dementia and Alzheimer disease. Ataxin-2 is one of many genes that, when altered, causes SCA. It normally carries a run of 22 or 23 glutamines; if that number goes above 34, SCA results. Now, Gitler and colleagues propose that 27-33 glutamines in a row predispose a person to ALS. “We are not saying that these expansions cause the disease,” he was careful to note. “We do see some of these expansions in controls.”
Gitler and colleagues exploited a yeast model for TDP-43 proteinopathy (see ARF related news story on Johnson et al., 2008). They screened 5,500 other yeast genes, from an overexpression library, for candidates that suppress or enhance TDP-43 toxicity. Of the 40 leads to come out of that screen, one immediately raised interest. The yeast Pab1-binding protein 1, or Pbp1, an ortholog of human ataxin-2, made TDP-43 toxicity worse when added to the gene pool. Furthermore, if the researchers deleted Pbp1 from the TDP-43 yeast, the yeast stayed healthy, suggesting Pbp1 is essential for TDP-43 toxicity.
Yeast, of course, lack a nervous system, so Gitler wanted to test the ataxin-2/TDP-43 interaction in a simple system that did. Collaborator Nancy Bonini had both a TDP-43 overexpression fly model that exhibits motor neuron degeneration and a Drosophila line in which the ataxin-2 gene was upregulated. Crossing the two, the researchers found that, as in yeast, the extra ataxin-2 increased disease severity in TDP-43-expressing flies. “This meant that this interaction was conserved in the nervous system,” Gitler said.
So what about mammals? The researchers examined TDP-43 and ataxin-2 in the context of human embryonic kidney cells, showing by co-immunoprecipitation that ataxin-2 and TDP-43 interact. However, that binding depended on RNA. When they mutated TDP-43’s RNA-binding domain, it no longer pulled down ataxin-2. The authors suggest that an unknown RNA molecule may bridge the two, or that binding to RNA may make TDP-43 able to also bind ataxin-2.
Finally, the researchers recruited Trojanowski and Lee to examine human tissue samples. They stained for ataxin-2 in spinal cord neurons from people who died of ALS and other causes. While the protein is normally diffuse in the cytoplasm, in ALS cases it formed abnormal cytoplasmic aggregates. These did not necessarily colocalize with TDP-43.
Gitler hypothesized that, since ataxin-2 is a polyglutamine-containing protein, the length of the polyglutamine region might be involved in ALS. So he collected DNA samples from 915 people with ALS and 980 healthy controls to measure the length of the ataxin-2 genes. “It was just me in the lab doing this experiment,” he recalled, “because everyone thought it was crazy.”
Gitler discovered that nearly 5 percent of ALS cases had 27-33 glutamine repeats in ataxin-2, compared to the normal 22 or 23. Less than 1 percent of control cases had as many, while glutamine repeats numbering above 31 existed only in people who had ALS.
The researchers propose that ataxin-2 is somehow required for TDP-43 toxicity. Perhaps, Gitler mused, the intermediate-length ataxin-2 is difficult for the cell to degrade, so it hangs around longer and promotes TDP-43 sequestration in the cytoplasm. “Identification of an association between ataxin-2 and ALS also provides additional evidence that altered RNA processing may be central to this disorder,” write Clotilde Lagier-Tourenne and Don Cleveland of the University of California in a commentary accompanying the paper.
If TDP-43 and ataxin-2 must meet in order to cause trouble, then interfering with the reaction could be a therapeutic strategy for ALS, Gitler suggested. However, it might be tricky to target a drug to only the intermediate-polyglutamine form of ataxin-2, noted Leonard Petrucelli of the Mayo Clinic in Jacksonville, Florida, who was not involved with the study.
The paper “adds an additional wrinkle” to the story scientists are decoding about TDP-43’s toxic activity, Petrucelli said. Recent animal models, which lack the noticeable cytoplasmic TDP-43 inclusions of the human disease but still get ill, had him and others thinking that the TDP-43 remaining in the nucleus was the true killer (see ARF related news story on Xu et al., 2010; ARF related news story on Wils et al., 2010; and ARF related news story on Wegorzewska et al., 2009). But since ataxin-2 is cytoplasmic, the current work suggests that is where the toxic events unfold.
In addition, Petrucelli was intrigued that the RNA-binding domain of TDP-43 was required for binding ataxin-2 and causing pathology. Work in his lab indicates that TDP-43 is cleaved in the cytoplasm, with carboxyl-terminal fragments wreaking the most havoc (see ARF related news story on Zhang et al., 2009). But the protein would have to be full length, or near to it, to be able to bind RNA, he said.
“Regardless of the species, regardless of whether it is full length or fragment, it is the cytoplasmic TDP-43 that is toxic,” Petrucelli concluded.
Although SCA and ALS are distinct diseases, there are occasional clinical cases that meld the two (reviewed in Nanetti et al., 2009). Trojanowski and colleagues examined spinal cord tissue from two people who had SCA caused by an ataxin-2 mutation. And they found mislocalized TDP-43. “Maybe these two diseases are not so different after all,” Gitler said. He and his coauthors suggest that ataxin-2 repeat disease constitutes a spectrum, with longer repeats leaning toward spinocerebellar pathology and mid-length stretches pushing toward disease the lower spinal motor neurons.
“This might just be the tip of the iceberg for ataxin-2, which could contribute to the pathogenesis of many other diseases involving TDP-43,” the authors write. They have already examined two frontotemporal lobar dementia cases, and discovered ataxin-2 inclusions lurking there along with TDP-43. Now, Gitler said, they are looking for evidence of ataxin-2 involvement in Parkinson disease and other polyglutamine conditions, and also examining other polyglutamine proteins in ALS. Gitler is also testing whether FUS—another RNA-binding protein that, when mutated, causes ALS—interacts with ataxin-2.—Amber Dance.
Elden AC, Kim HJ, Hart MP, Chen-Plotkin A, Johnson BS, Fang X, Armakola M, Geser F, Greene R, Lu MM, Padmanabhan A, Clay-Falcone D, McCluskey L, Elman L, Juhr D, Gruber PJ, Rüb U, Auburger G, Trojanowski JQ, Lee VMY, Van Deerlin VM, Bonini NM, Gitler AD. Ataxin-2 intermediate-length polyglutamine expansions are associated with increased risk for ALS. Nature. 2010 Aug 26;466:1069-75. Abstract
Lagier-Tourenne C, Cleveland DW. An expansion in ALS genetics. Nature. 2010 Aug 26;466:1052-3. Abstract